Drug-likeness analysis of traditional Chinese medicines: 1. property distributions of drug-like compounds, non-drug-like compounds and natural compounds from traditional Chinese medicines

Abstract

BackgroundIn this work, we analyzed and compared the distribution profiles of a wide variety of molecular properties for three compound classes: drug-like compounds in MDL Drug Data Report (MDDR), non-drug-like compounds in Available Chemical Directory (ACD), and natural compounds in Traditional Chinese Medicine Compound Database (TCMCD).ResultsThe comparison of the property distributions suggests that, when all compounds in MDDR, ACD and TCMCD with molecular weight lower than 600 were used, MDDR and ACD are substantially different while TCMCD is much more similar to MDDR than ACD. However, when the three subsets of ACD, MDDR and TCMCD with similar molecular weight distributions were examined, the distribution profiles of the representative physicochemical properties for MDDR and ACD do not differ significantly anymore, suggesting that after the dependence of molecular weight is removed drug-like and non-drug-like molecules cannot be effectively distinguished by simple property-based filters; however, the distribution profiles of several physicochemical properties for TCMCD are obviously different from those for MDDR and ACD. Then, the performance of each molecular property on predicting drug-likeness was evaluated. No single molecular property shows good performance to discriminate between drug-like and non-drug-like molecules. Compared with the other descriptors, fractional negative accessible surface area (FASA-) performs the best. Finally, a PCA-based scheme was used to visually characterize the spatial distributions of the three classes of compounds with similar molecular weight distributions.ConclusionIf FASA- was used as a drug-likeness filter, more than 80% molecules in TCMCD were predicted to be drug-like. Moreover, the principal component plots show that natural compounds in TCMCD have different and even more diverse distributions than either drug-like compounds in MDDR or non-drug-like compounds in ACD.

The application of in silico drug-likeness predictions in pharmaceutical research

“…For example, our studies showed that Rule-of-Five is not a good predictor to estimate intestinal absorption, and even less reliable than a simple molecular property, topological polar surface area (TPSA) [58,59]. In 2012, we evaluated the performances of Rule-of-Five and Opera's filters to distinguish the drug-like molecules in MDDR from the non-drug-like molecules in ACD [50]. Two MDDR and ACD subsets with similar MW distributions were generated artificially.…”

“…That is to say, when drug-like and non-drug-like compounds have similar MW distributions, Rule-of-Five does not have any prediction capability to distinguish drug-like from non-drug-like molecules. We also applied the Opera's RNG and RGB filters to evaluate the ACD and MDDR subsets with similar MW distributions, and most non-drug-like molecules in ACD (59.5%) were identified to be drug-like [50]. Recently, Ritchie et al evaluated the correlations between the QED score and the actual pharmaceutical and pharmacokinetic (PK) profiles in humans for 300 oral drugs.…”

“…Recently, we analyzed the distribution profiles of the physicochemical properties for the drug-like molecules in MDDR and the non-drug-like molecules in ACD [50], and found that the distributions of numerous important molecular physicochemical properties for drug-like and non-drug-like molecules have intrinsic differences. However, by removing the dependence of A C C E P T E D M A N U S C R I P T ACCEPTED MANUSCRIPT 6 molecular weight, unlike the Opera's results [24], the property distributions of drug-like and non-drug-like molecules do not differ significantly anymore (Figure 1), illustrating that a single physicochemical property may not be an effective filter to distinguish drug-like from non-drug-like molecules.…”

Fragment virtual screening based on Bayesian categorization for discovering novel VEGFR-2 scaffolds

“…macrofocus.com/) were used to analyze the structural diversity. Scaffold architectures were first clustered using ECFP_4 fingerprints by the cluster molecules component in Pipeline Pilot with the average number of compounds per cluster was set to 50 [32]. Then, the scaffold cluster number as well as its frequency were utilized to categorize scaffolds by circles of different area and colors [31].…”

“…The chemical space diversity of all the databases was evaluated to ensure an adequate novelty not only from the 2D physiochemical properties aspect but also from the scaffold and structural diversity perspectives. The latter were explicated through the count of scaffolds, cumulative scaffold frequency plot (CSFP), P n values as well as tree maps visualization [31,32]. Simultaneously, Bayesian models were constructed using VEGFR-2 actives and decoys from the directory of useful decoys enhanced database (DUD-E) [33] database and their reliability was validated using internal [16,19,22].…”